Abstract:

A cross-response liquid array sensor that simulated the mammalian taste system was developed for the determination of liquor aromatic types. The liquid array sensor was constructed with 8 sensing units which were consisted of 7 chemical dyes and 1 porphyrin compound. It achieved the type identification through the analyte fingerprints constituted by congregating the spectral response signal of 8 sensing units. The data were collected via a 96-well plate microplate reader, and then was further analyzed with principal component analysis(PCA), hierarchical clustering analysis(HCA) and linear discriminant analysis(LDA). We detected 9 different aromatic liquors. PCA results suggested that the mechanism of identification was mainly based on the trace components of Chinese liquor. Thereamong, the acids of Chinese liquor were the maximum contribution for the identification, because the contribution rate of acids was 54.3% and that of aromatic substances was only 18.6%. Simultaneously, only by 63.4% data rate could distinct liquor aromatic types with satisfaction. HCA dendrogram demonstrated that 9 different aromatic liquors could be correctly classified. LDA further showed the classification accuracy of identification for these 9 different aromatic liquors was 100%. These results showed that this method may provide a simple and efficient way for the rapid identification of liquor.

Key words: Liquor aromatic types, Taste biomimetic, Liquid array sensor, Cross-reactive